library(dplyr)
## 
## Attaching package: 'dplyr'
## The following objects are masked from 'package:stats':
## 
##     filter, lag
## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union
library(ggthemes)
library(ggplot2)
library(e1071)
library(rpivotTable)
library(foreign)
library(readxl)
library(plotly)
## 
## Attaching package: 'plotly'
## The following object is masked from 'package:ggplot2':
## 
##     last_plot
## The following object is masked from 'package:stats':
## 
##     filter
## The following object is masked from 'package:graphics':
## 
##     layout
library(readr)
library(purrr)
library(gridExtra)
## 
## Attaching package: 'gridExtra'
## The following object is masked from 'package:dplyr':
## 
##     combine
library(tidyr)
library(psych)
## 
## Attaching package: 'psych'
## The following objects are masked from 'package:ggplot2':
## 
##     %+%, alpha
library(HSAUR2)
## Loading required package: tools
library(lubridate)
## 
## Attaching package: 'lubridate'
## The following object is masked from 'package:base':
## 
##     date

Importing the Data Sets

yr_12_15 = read.csv('C:/Users/Halappa Guruppanavar/Desktop/Github Upload Documents/DataSets/Global Terrorist Data/gtd_12to15_52134.csv')
yr_70_91 = read.csv('C:/Users/Halappa Guruppanavar/Desktop/Github Upload Documents/DataSets/Global Terrorist Data/gtd_70to91_49566.csv')
yr_92_11 = read.csv('C:/Users/Halappa Guruppanavar/Desktop/Github Upload Documents/DataSets/Global Terrorist Data/gtd_92to11_no 93_55072.csv')
yr_93 = read.csv('C:/Users/Halappa Guruppanavar/Desktop/Github Upload Documents/DataSets/Global Terrorist Data/gtd1993_748.csv')
  1. Number of attacks per year.
df1 = yr_70_91 %>% select(iyear)
df2 = yr_92_11 %>% select(iyear)
df3 = yr_93 %>% select(iyear)
df4 = yr_12_15 %>% select(iyear)

no_of_attacks = rbind(df1,df2,df3,df4) %>% group_by(iyear) %>% summarise(total_attacks = n())

## ggplot
scale_x_continuous(breaks = c(1970:2015))
## <ScaleContinuousPosition>
##  Range:  
##  Limits:    0 --    1
area = ggplot(no_of_attacks, aes(iyear, total_attacks)) + scale_x_continuous(breaks = c(1970:2015)) +
  theme_dark() + theme(axis.text.x = element_text(angle = 45, vjust = .6)) + geom_point() +
  geom_area(size = .8, color = 'black', fill = 'tomato2', alpha = .8) +
  labs(title = 'Total Number of attacks per year',
       x = 'Year',
       y = 'Attacks')
#area
ggplotly(area)
  1. Number of bombings per year.
df2_1 = yr_70_91 %>% filter(attacktype1_txt == 'Bombing/Explosion') %>% select(iyear)
df2_2 = yr_92_11 %>% filter(attacktype1_txt == 'Bombing/Explosion') %>% select(iyear)
df2_3 = yr_93 %>% filter(attacktype1_txt == 'Bombing/Explosion') %>% select(iyear)
df2_4 = yr_12_15 %>% filter(attacktype1_txt == 'Bombing/Explosion') %>% select(iyear)

total_bomb_py = rbind(df2_1,df2_2,df2_3,df2_4) %>% group_by(iyear) %>% summarise(tot_bom = n())

comb = merge(x= no_of_attacks,
             y = total_bomb_py,
             by.x = 'iyear',
             by.y = 'iyear',
             all.x = TRUE)

#comb$iyear = as.Date(comb$iyear, '%Y') #converting char year to typical year

stack_area = ggplot(comb, aes(iyear)) +  
#  scale_x_continuous(breaks = c(1970:2015)) +
  geom_area(aes(y = total_attacks, col = 'Total Attacks'), fill='lightsalmon3', size= 0.8, alpha = .8) +
  geom_point(aes(y = total_attacks)) +
  geom_area(aes(y = tot_bom, col = 'Total bombings'), fill = 'aquamarine3', size = .8, alpha = .6) + 
  geom_point(aes(y = tot_bom)) +
  theme_economist_white() + 
  theme(axis.text.x = element_text(angle = 0, vjust = .6),
        legend.position = 'top') +
  labs(title = 'Comparision between Number of Attacks and Bombings per year',
       x = 'Year',
       y = ' ')
stack_area

  1. Terrorist attacks region wise per year.
df3_1 = yr_70_91 %>% group_by(iyear, region_txt) %>% summarise(total_attacks = n())
df3_2 = yr_92_11 %>% group_by(iyear, region_txt) %>% summarise(total_attacks = n())
df3_3 = yr_93 %>% group_by(iyear, region_txt) %>% summarise(total_attacks = n())
df3_4 = yr_12_15 %>% group_by(iyear, region_txt) %>% summarise(total_attacks = n())

reg_wise_att = rbind(df3_1,df3_2,df3_3,df3_4)
## Warning in bind_rows_(x, .id): Unequal factor levels: coercing to character
## Warning in bind_rows_(x, .id): binding character and factor vector,
## coercing into character vector

## Warning in bind_rows_(x, .id): binding character and factor vector,
## coercing into character vector

## Warning in bind_rows_(x, .id): binding character and factor vector,
## coercing into character vector
reg_wise_att$iyear = as.factor(reg_wise_att$iyear)

reg_wise = ggplot(reg_wise_att , aes(x=region_txt, y=iyear)) + 
  geom_point(aes(col=region_txt, size=total_attacks)) + theme_gray() +
  theme(axis.text.x = element_blank()) +
  labs(y="Year", 
       x="Region", 
       title="Region-wise attacks per year")
reg_wise

  1. Top 5 type of terror attacks per region.
yr_93$region_txt[394] = 'Eastern Europe'
yr_93$region_txt[412] = 'Eastern Europe'
yr_93$region_txt[507] = 'Eastern Europe'
yr_93$region_txt[733] = 'Eastern Europe'
yr_93$region_txt[423] = 'Eastern Europe'
yr_93$region_txt[429] = 'Eastern Europe'
yr_93$region_txt[514] = 'Eastern Europe'
yr_93$region_txt[536] = 'Eastern Europe'
yr_93$region_txt[538] = 'Eastern Europe'

d4_1 = yr_70_91 %>% group_by(region_txt, attacktype1_txt) %>% summarise(total_attacks = n())
d4_2 = yr_92_11 %>% group_by(region_txt, attacktype1_txt) %>% summarise(total_attacks = n())
d4_3 = yr_93 %>% group_by(region_txt, attacktype1_txt) %>% summarise(total_attacks = n())
d4_4 = yr_12_15 %>% group_by(region_txt, attacktype1_txt) %>% summarise(total_attacks = n())


reg_wise_att_t5 = rbind(d4_1,d4_2,d4_3,d4_4) %>% 
  group_by(region_txt, attacktype1_txt) %>% 
  summarise(total_attacks = sum(total_attacks)) %>% top_n(5)
## Warning in bind_rows_(x, .id): Unequal factor levels: coercing to character
## Warning in bind_rows_(x, .id): binding character and factor vector,
## coercing into character vector

## Warning in bind_rows_(x, .id): binding character and factor vector,
## coercing into character vector

## Warning in bind_rows_(x, .id): binding character and factor vector,
## coercing into character vector
## Selecting by total_attacks
top.5.att.r.w = ggplot(reg_wise_att_t5, aes(attacktype1_txt, total_attacks)) +
  geom_bar(stat = 'Identity', width = .5, aes(fill = attacktype1_txt))+
  theme_dark()+
  theme(axis.text.x = element_blank(),
        axis.title = element_blank(),
        axis.ticks.x = element_blank(),
        axis.text.y = element_text(angle = 0),
        legend.title = element_blank(),
        legend.position = 'bottom') +
  facet_wrap(~region_txt)

top.5.att.r.w

  1. Heaviest hit Target types (Based on both Killed and wounded).
df5_1 = yr_70_91 %>% group_by(targtype1_txt) %>% summarise(kills = sum(nkill, na.rm = TRUE),
                                                           wound = sum(nwound, na.rm = TRUE))
df5_2 = yr_92_11 %>% group_by(targtype1_txt) %>% summarise(kills = sum(nkill, na.rm = TRUE),
                                                           wound = sum(nwound, na.rm = TRUE))
df5_3 = yr_93 %>% group_by(targtype1_txt) %>% summarise(kills = sum(nkill, na.rm = TRUE),
                                                           wound = sum(nwound, na.rm = TRUE))
df5_4 = yr_12_15 %>% group_by(targtype1_txt) %>% summarise(kills = sum(nkill, na.rm = TRUE),
                                                           wound = sum(nwound, na.rm = TRUE))

hvy_hit = rbind(df5_1, df5_2, df5_3, df5_4) %>% group_by(targtype1_txt) %>%
  summarise(kills = round(sum(kills),0),
            wound = round(sum(wound),0),
            total = kills + wound) %>% arrange(-total) %>% head(10)

hvy_hit_pie = plot_ly(hvy_hit, labels = ~targtype1_txt, values = ~total, 
                      type = 'pie', pull = .01, textposition = 'inside') %>%
  layout(title = 'HEAVIEST HIT TARGET TYPE.',
         xaxis = list(showgrid = FALSE, zeroline = FALSE, showticklabels = FALSE),
         yaxis = list(showgrid = FALSE, zeroline = FALSE, showticklabels = FALSE))

hvy_hit_pie
  1. Terrorist attack in India and Pakistan in last 45 years.
#str(yr_70_91)
df6_1 = yr_70_91 %>% filter(country_txt %in% c('Pakistan','India'))%>%
  group_by(iyear ,country_txt) %>%
  summarise(total_attacks = n())

df6_2 = yr_92_11 %>% filter(country_txt %in% c('Pakistan','India'))%>%
  group_by(iyear, country_txt) %>%
  summarise(total_attacks = n())

df6_3 = yr_93 %>% filter(country_txt %in% c('Pakistan','India'))%>%
  group_by(iyear, country_txt) %>%
  summarise(total_attacks = n())

df6_4 = yr_12_15 %>% filter(country_txt %in% c('Pakistan','India'))%>%
  group_by(iyear, country_txt) %>%
  summarise(total_attacks = n())

ind_pak = rbind(df6_1,df6_2,df6_3,df6_4) %>% group_by(iyear, country_txt) %>%
  summarise(total_attacks = sum(total_attacks))
## Warning in bind_rows_(x, .id): Unequal factor levels: coercing to character
## Warning in bind_rows_(x, .id): binding character and factor vector,
## coercing into character vector

## Warning in bind_rows_(x, .id): binding character and factor vector,
## coercing into character vector

## Warning in bind_rows_(x, .id): binding character and factor vector,
## coercing into character vector

## Warning in bind_rows_(x, .id): binding character and factor vector,
## coercing into character vector
ivp_comp = ggplot(ind_pak, aes(iyear, total_attacks)) +
  geom_area(aes(fill = country_txt), size = 1.2, 
            color = 'black',alpha = .8)+ 
  geom_point(size = 2) +
  theme_grey()+
  theme(legend.title = element_blank(), 
        legend.position = 'none') +
  labs(title = 'INDIA vs. PAKISTAN',
       x = 'Years',
       y = 'Attacks') +
  facet_wrap(~country_txt)

ggplotly(ivp_comp)
  1. Terror attack in United States vs Russian Federation/USSR in last 45 years.
df7_1 = yr_70_91 %>% filter(country_txt %in% c('United States', 'Soviet Union')) %>%
  group_by(iyear, country_txt) %>%
  summarise(total_attacks = n())

df7_2 = yr_92_11 %>% filter(country_txt %in% c('United States', 'Russia')) %>%
  group_by(iyear, country_txt) %>%
  summarise(total_attacks = n())

df7_3 = yr_93 %>% filter(country_txt %in% c('United States', 'Russia')) %>%
  group_by(iyear, country_txt) %>%
  summarise(total_attacks = n())

df7_4 = yr_12_15 %>% filter(country_txt %in% c('United States', 'Russia'))%>%
  group_by(iyear, country_txt) %>%
  summarise(total_attacks = n())


df7_5 = rbind(df7_1,df7_2,df7_3,df7_4) %>%
  group_by(iyear, country_txt) %>% summarise(total_attacks = sum(total_attacks))
## Warning in bind_rows_(x, .id): Unequal factor levels: coercing to character
## Warning in bind_rows_(x, .id): binding character and factor vector,
## coercing into character vector

## Warning in bind_rows_(x, .id): binding character and factor vector,
## coercing into character vector

## Warning in bind_rows_(x, .id): binding character and factor vector,
## coercing into character vector

## Warning in bind_rows_(x, .id): binding character and factor vector,
## coercing into character vector
df7_5$country_txt[9] = 'Russia' # to change soviet union into russia
df7_5$country_txt[21] = 'Russia'
df7_5$country_txt[23] = 'Russia'
df7_5$country_txt[25] = 'Russia'

df7_6 = df7_5 %>% group_by(iyear, country_txt) %>% 
  summarise(total_attacks = sum(total_attacks)) # adding up to the total

usvr_comp = ggplot(df7_6, aes(iyear, total_attacks)) +
  geom_area(aes(fill = country_txt), size = 1.2, 
            color = 'black',alpha = .8)+ 
  geom_point(size = 2) +
  theme_economist_white()+
  theme(legend.title = element_blank(), 
        legend.position = 'none') +
  labs(title = 'United States vs Russia.',
       x = 'Years',
       y = 'Attacks') +
  facet_wrap(~country_txt)

ggplotly(usvr_comp)
  1. Where are there the most casualties?.
df8_1 = yr_70_91 %>% group_by(country_txt) %>% summarise(kills = sum(nkill, na.rm = TRUE),
                                                           wound = sum(nwound, na.rm = TRUE))
df8_2 = yr_92_11 %>% group_by(country_txt) %>% summarise(kills = sum(nkill, na.rm = TRUE),
                                                           wound = sum(nwound, na.rm = TRUE))
df8_3 = yr_93 %>% group_by(country_txt) %>% summarise(kills = sum(nkill, na.rm = TRUE),
                                                        wound = sum(nwound, na.rm = TRUE))
df8_4 = yr_12_15 %>% group_by(country_txt) %>% summarise(kills = sum(nkill, na.rm = TRUE),
                                                           wound = sum(nwound, na.rm = TRUE))

most_casualties = rbind(df8_1,df8_2,df8_3,df8_4) %>% group_by(country_txt) %>% 
  summarise(kills = round(sum(kills),0),
            wound = round(sum(wound),0),
            total_casualties = kills + wound) %>% arrange(-total_casualties) %>%
  head(15)

cas_bar = ggplot(most_casualties, 
                 aes(x = reorder(country_txt, -total_casualties), y = total_casualties)) +
  geom_bar(stat = 'Identity', width = 0.7, aes(fill = country_txt)) + theme_economist() +
  theme(legend.title = element_blank(),
        legend.position = 'none')+
  labs(title = 'Most affected Countries.',
       x = 'Country',
       y = 'Casualties')
ggplotly(cas_bar)
  1. How have casualties evolved throughout the years?
df9_1 = yr_70_91 %>% group_by(iyear) %>% summarise(kills = sum(nkill, na.rm = TRUE),
                                                         wound = sum(nwound, na.rm = TRUE))
df9_2 = yr_92_11 %>% group_by(iyear) %>% summarise(kills = sum(nkill, na.rm = TRUE),
                                                         wound = sum(nwound, na.rm = TRUE))
df9_3 = yr_93 %>% group_by(iyear) %>% summarise(kills = sum(nkill, na.rm = TRUE),
                                                      wound = sum(nwound, na.rm = TRUE))
df9_4 = yr_12_15 %>% group_by(iyear) %>% summarise(kills = sum(nkill, na.rm = TRUE),
                                                         wound = sum(nwound, na.rm = TRUE))

casualties_by_yr = rbind(df9_1,df9_2,df9_3,df9_4) %>% group_by(iyear) %>% 
  summarise(kills = round(sum(kills),0),
            wound = round(sum(wound),0),
            total_casualties = kills + wound)


cas_plot =  ggplot(casualties_by_yr, aes(iyear)) + theme_solarized_2() + 
#  scale_x_continuous(breaks = c(1970:2015)) +
  geom_line(aes(y = kills, col = 'Kills'), size = 1.2) + 
  geom_point(aes(y = kills), color = 'red')+
  geom_line(aes(y = wound, col = 'Wound'), size = 1.2) + 
  geom_point(aes(y = wound), color = 'blue') +
  theme(axis.text.x = element_text(angle = 0, vjust = .6),
        legend.title = element_blank()) +
  labs(title = 'Evolution of Casualties',
       x = 'Year',
       y = 'Casualties')
ggplotly(cas_plot)
  1. What are the casualties by weapon type?
df10_1 = yr_70_91 %>% group_by(weaptype1_txt) %>% summarise(kills = sum(nkill, na.rm = TRUE),
                                                         wound = sum(nwound, na.rm = TRUE))
df10_2 = yr_92_11 %>% group_by(weaptype1_txt) %>% summarise(kills = sum(nkill, na.rm = TRUE),
                                                         wound = sum(nwound, na.rm = TRUE))
df10_3 = yr_93 %>% group_by(weaptype1_txt) %>% summarise(kills = sum(nkill, na.rm = TRUE),
                                                      wound = sum(nwound, na.rm = TRUE))
df10_4 = yr_12_15 %>% group_by(weaptype1_txt) %>% summarise(kills = sum(nkill, na.rm = TRUE),
                                                         wound = sum(nwound, na.rm = TRUE))

casualties_by_weap = rbind(df10_1,df10_2,df10_3,df10_4) %>% group_by(weaptype1_txt) %>% 
  summarise(kills = round(sum(kills),0),
            wound = round(sum(wound),0),
            total_casualties = kills + wound)

casualties_by_weap$weaptype1_txt = as.character(casualties_by_weap$weaptype1_txt)
casualties_by_weap$weaptype1_txt[12] = 'Vehicles'

weap_cas_plot =  plot_ly(casualties_by_weap, labels = ~weaptype1_txt, values = ~total_casualties, 
        type = 'pie', pull = .01, textposition = 'inside',
        textinfo = 'value+percent') %>%
  layout(title = 'CASUALTIES BY WEAPON TYPE.',
         xaxis = list(showgrid = FALSE, zeroline = FALSE, showticklabels = FALSE),
         yaxis = list(showgrid = FALSE, zeroline = FALSE, showticklabels = FALSE))

weap_cas_plot
  1. Are certain nationalities more targeted? If yes, which one?
df11_1 = yr_70_91 %>% group_by(natlty1_txt) %>% summarise(kills = sum(nkill, na.rm = TRUE),
                                                            wound = sum(nwound, na.rm = TRUE))
df11_2 = yr_92_11 %>% group_by(natlty1_txt) %>% summarise(kills = sum(nkill, na.rm = TRUE),
                                                            wound = sum(nwound, na.rm = TRUE))
df11_3 = yr_93 %>% group_by(natlty1_txt) %>% summarise(kills = sum(nkill, na.rm = TRUE),
                                                         wound = sum(nwound, na.rm = TRUE))
df11_4 = yr_12_15 %>% group_by(natlty1_txt) %>% summarise(kills = sum(nkill, na.rm = TRUE),
                                                            wound = sum(nwound, na.rm = TRUE))

nationalities = rbind(df11_1,df11_2,df11_3,df11_4) %>% group_by(natlty1_txt) %>% 
  summarise(kills = round(sum(kills),0),
            wound = round(sum(wound),0),
            total_casualties = kills + wound) %>% arrange(-total_casualties)%>%
  head(10)

nl = gather(nationalities, cas_type, casualties, kills:wound)

natlt =  ggplot(nl, aes(fill=cas_type, y=casualties, x=reorder(natlty1_txt, -casualties))) +
  theme_fivethirtyeight() +
  geom_bar(position="dodge", stat="identity", width = .5) +
  theme(legend.title = element_blank(),
        legend.position = c(.9,.9)) +
  labs(title = 'Most targeted nationalities.',
       x = 'Nationalities',
       y = 'Casualties')

ggplotly(natlt)
## Warning: plotly.js does not (yet) support horizontal legend items 
## You can track progress here: 
## https://github.com/plotly/plotly.js/issues/53
  1. Countries better at defending themselves.
df12_1 = yr_70_91 %>% group_by(country_txt) %>% summarise(failure = sum(success == 0, na.rm = TRUE),
                                                          tot = n(),
                                                          def_rate = failure/tot)
df12_2 = yr_92_11 %>% group_by(country_txt) %>%summarise(failure = sum(success == 0, na.rm = TRUE),
                                                         tot = n(),
                                                         def_rate = failure/tot)
df12_3 = yr_93 %>% group_by(country_txt) %>% summarise(failure = sum(success == 0, na.rm = TRUE),
                                                       tot = n(),
                                                       def_rate = failure/tot)
df12_4 = yr_12_15 %>% group_by(country_txt) %>%summarise(failure = sum(success == 0, na.rm = TRUE),
                                                         tot = n(),
                                                         def_rate = failure/tot)

# for more than 250 attacks
t1 = rbind(df12_1,df12_2,df12_3,df12_4) %>% group_by(country_txt) %>% 
  summarise(failure = sum(failure),
            total = sum(tot),
            def_rate = round(failure*100/total),0) %>% filter(total>250) %>% 
  arrange(-def_rate) %>% head(10)

str_def_t1 =  ggplot(t1, aes(x = reorder(country_txt, -def_rate), y = def_rate)) +
  geom_bar(stat = 'Identity', width = 0.5, aes(fill = country_txt)) +
  geom_text(aes(label = def_rate),
            position = position_dodge(width = 0.9),
            vjust = -0.2) +
  theme_economist() +
  theme(legend.title = element_blank(),
        legend.position = 'none')+
  labs(title = 'Countries with strong defence.',
       subtitle = 'For more than 250 attacks.',
       x = 'Country',
       y = 'Defence rate per 100 attacks')

str_def_t1